In the case where the target timestamp is before the origin timestamp
and their difference is already an exact multiple of the stride, the
code incorrectly subtracted the stride anyway.
Also detect several integer-overflow cases that previously produced
bogus results. (The submitted patch tried to avoid overflow, but
I'm not convinced it's right, and problematic cases are so far out of
the plausibly-useful range that they don't seem worth sweating over.
Let's just use overflow-detecting arithmetic and throw errors.)
timestamp_bin() and timestamptz_bin() are basically identical and
so had identical bugs. Fix both.
Report and patch by Moaaz Assali, adjusted some by me. Back-patch
to v14 where date_bin() was introduced.
Discussion: https://postgr.es/m/CALkF+nvtuas-2kydG-WfofbRSJpyODAJWun==W-yO5j2R4meqA@mail.gmail.com
This commit fixes the intermittent buildfarm failures in 031_column_list.
I missed to back-patch while committing b6df0798a5 in the HEAD.
Diagnosed-by: Amit Kapila
Author: Vignesh C
Backpatch-through: 15
Discussion: https://postgr.es/m/3307255.1706911634@sss.pgh.pa.us
When this assertion was installed (in commit d2f60a3ab), I thought
it was only for catching server logic errors that caused accesses to
catalogs that were undergoing index rebuilds. However, it will also
fire in case of a user-defined index expression that attempts to
access its own table. We occasionally see reports of people trying
to do that, and typically getting unintelligible low-level errors
as a result. We can provide a more on-point message by making this
a regular runtime check.
While at it, adjust the similar error check in
systable_beginscan_ordered to use the same message text. That one
is (probably) not reachable without a coding bug, but we might as
well use a translatable message if we have one.
Per bug #18363 from Alexander Lakhin. Back-patch to all supported
branches.
Discussion: https://postgr.es/m/18363-e3598a5a572d0699@postgresql.org
build_child_join_sjinfo creates a derived SpecialJoinInfo in
the short-lived GEQO context, but afterwards the semi_rhs_exprs
from that may be used in a UniquePath for a child base relation.
This breaks the expectation that all base-relation-level structures
are in the planning-lifespan context, leading to use of a dangling
pointer with probable ensuing crash later on in create_unique_plan.
To fix, copy the expression trees when making a UniquePath.
Per bug #18360 from Alexander Lakhin. This has been broken since
partitionwise joins were added, so back-patch to all supported
branches.
Discussion: https://postgr.es/m/18360-a23caf3157f34e62@postgresql.org
Verify that a user running MERGE with a DO NOTHING clause has
privileges to read the table, even if no columns are referenced. Such
privileges were already required if the ON clause or any of the WHEN
conditions referenced any column at all, so there's no functional change
in practice.
This change fixes an assertion failure in the case where no column is
referenced by the command and the WHEN clauses are all DO NOTHING.
Backpatch to 15, where MERGE was introduced.
Reported-by: Alena Rybakina <a.rybakina@postgrespro.ru>
Reported-by: Alexander Lakhin <exclusion@gmail.com>
Discussion: https://postgr.es/m/4d65a385-7efa-4436-a825-0869f89d9d92@postgrespro.ru
Partition pruning wrongly assumed that, for a table partitioned on a
boolean column, a clause in the form "boolcol IS NOT false" and "boolcol
IS NOT true" could be inverted to correspondingly become "boolcol IS true"
and "boolcol IS false". These are not equivalent as the NOT version
matches the opposite boolean value *and* NULLs. This incorrect assumption
meant that partition pruning pruned away partitions that could contain
NULL values.
Here we fix this by correctly not pruning partitions which could store
NULLs.
To be affected by this, the table must be partitioned by a NULLable boolean
column and queries would have to contain "boolcol IS NOT false" or "boolcol
IS NOT true". This could result in queries filtering out NULL values
with a LIST partitioned table and "ERROR: invalid strategy number 0"
for RANGE and HASH partitioned tables.
Reported-by: Alexander Lakhin
Bug: #18344
Discussion: https://postgr.es/m/18344-8d3f00bada6d09c6@postgresql.org
Backpatch-through: 12
intoasc(), a wrapper for PGTYPESinterval_to_asc that converts an
interval to its textual representation, used a plain memcpy() when
copying its result. This could miss a zero-termination in the result
string, leading to an incorrect result.
The routines in informix.c do not provide the length of their result
buffer, which would allow a replacement of strcpy() to safer strlcpy()
calls, but this requires an ABI breakage and that cannot happen in
back-branches.
Author: Oleg Tselebrovskiy
Reviewed-by: Ashutosh Bapat
Discussion: https://postgr.es/m/bf47888585149f83b276861a1662f7e4@postgrespro.ru
Backpatch-through: 12
The macOS Finder application creates .DS_Store files in directories
when opened, which creates problems for serverside utilities which
expect all files to be PostgreSQL specific files. Skip these files
when encountered in pg_checksums, pg_rewind and pg_basebackup.
This was extracted from a larger patchset for skipping hidden files
and system files, where the concencus was to just skip these. Since
this is equally likely to happen in every version, backpatch to all
supported versions.
Reported-by: Mark Guertin <markguertin@gmail.com>
Reviewed-by: Michael Paquier <michael@paquier.xyz>
Reviewed-by: Tobias Bussmann <t.bussmann@gmx.net>
Discussion: https://postgr.es/m/E258CE50-AB0E-455D-8AAD-BB4FE8F882FB@gmail.com
Backpatch-through: v12
Since its introduction, pg_get_expr() has intended to silently
return NULL if called with an invalid relation OID, as can happen
when scanning the catalogs concurrently with relation drops.
However, there is a race condition: we check validity of the OID
at the start, but it could get dropped just afterward, leading to
failures. This is the cause of some intermittent instability we're
seeing in a proposed new test case, and presumably it's a hazard in
the field as well.
We can fix this by AccessShareLock-ing the target relation for the
duration of pg_get_expr(). Since we don't require any permissions
on the target relation, this is semantically a bit undesirable. But
it turns out that the set_relation_column_names() subroutine already
takes a transient AccessShareLock on that relation, and has done since
commit 2ffa740be in 2012. Given the lack of complaints about that, it
seems like there should be no harm in holding the lock a bit longer.
Back-patch to all supported branches.
Discussion: https://postgr.es/m/31ddcc01-a71b-4e8c-9948-01d1c47293ca@eisentraut.org
We previously supposed that it was okay for different threads to
call bindtextdomain() concurrently (cf. commit 1f655fdc3).
It now emerges that there's at least one gettext implementation
in which that triggers an abort() crash, so let's stop doing that.
Add mutexes guarding libpq's and ecpglib's calls, which are the
only ones that need worry about multithreaded callers.
Note: in libpq, we could perhaps have piggybacked on
default_threadlock() to avoid defining a new mutex variable.
I judge that not terribly safe though, since libpq_gettext could
be called from code that is holding the default mutex. If that
were the first such call in the process, it'd fail. An extra
mutex is cheap insurance against unforeseen interactions.
Per bug #18312 from Christian Maurer. Back-patch to all
supported versions.
Discussion: https://postgr.es/m/18312-bbbabc8113592b78@postgresql.org
Discussion: https://postgr.es/m/264860.1707163416@sss.pgh.pa.us
Fix pthread-win32.h and pthread-win32.c to provide a more complete
emulation of POSIX pthread mutexes: define PTHREAD_MUTEX_INITIALIZER
and make sure that pthread_mutex_lock() can operate on a mutex
object that's been initialized that way. Then we don't need the
duplicative platform-specific logic in default_threadlock() and
pgtls_init(), which we'd otherwise need yet a third copy of for
an upcoming bug fix.
Also, since default_threadlock() supposes that pthread_mutex_lock()
cannot fail, try to ensure that that's actually true, by getting
rid of the malloc call that was formerly involved in initializing
an emulated mutex. We can define an extra state for the spinlock
field instead.
Also, replace the similar code in ecpglib/misc.c with this version.
While ecpglib's version at least had a POSIX-compliant API, it
also had the potential of failing during mutex init (but here,
because of CreateMutex failure rather than malloc failure). Since
all of misc.c's callers ignore failures, it seems like a wise idea
to avoid failures here too.
A further improvement in this area could be to unify libpq's and
ecpglib's implementations into a src/port/pthread-win32.c file.
But that doesn't seem like a bug fix, so I'll desist for now.
In preparation for the aforementioned bug fix, back-patch to all
supported branches.
Discussion: https://postgr.es/m/264860.1707163416@sss.pgh.pa.us
The TransactionIdInRecentPast() should return false for all the transactions
older than TransamVariables->oldestClogXid. However, the function contains
a bug in comparison FullTransactionId to TransactionID allowing full
transactions between nextXid - 2^32 and oldestClogXid - 2^31.
This commit fixes TransactionIdInRecentPast() by turning the oldestClogXid into
FullTransactionId first, then performing the comparison.
Backpatch to all supported versions.
Reported-by: Egor Chindyaskin
Bug: 18212
Discussion: https://postgr.es/m/18212-547307f8adf57262%40postgresql.org
Author: Karina Litskevich
Reviewed-by: Kyotaro Horiguchi
Backpatch-through: 12
Fix for 344d62fb9a9: That commit introduced unlogged sequences and
made it so that identity/serial sequences automatically get the
persistence level of their owning table. But this works only for
CREATE TABLE and not for ALTER TABLE / ADD COLUMN. The latter would
always create the sequence as logged (default), independent of the
persistence setting of the table. This is fixed here.
Note: It is allowed to change the persistence of identity sequences
directly using ALTER SEQUENCE. So mistakes in existing databases can
be fixed manually.
Reviewed-by: Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>
Discussion: https://www.postgresql.org/message-id/flat/c4b6e2ed-bcdf-4ea7-965f-e49761094827%40eisentraut.org
When assertions are disabled, the built SQL statement is invalid and
you get a "syntax error". So this isn't a serious problem, but let's
avoid the assertion failure.
Backpatch to all supported versions.
Reviewed-by: Noah Misch
The internal commands in REFRESH MATERIALIZED VIEW CONCURRENTLY are
correctly executed in SECURITY_RESTRICTED_OPERATION mode, except for
creating the temporary "diff" table, because you cannot create
temporary tables in SRO mode. But creating the temporary "diff" table
is a pretty complex CTAS command that selects from another temporary
table created earlier in the command. If you can cajole that CTAS
command to execute code defined by the table owner, the table owner
can run code with the privileges of the user running the REFRESH
command.
The proof-of-concept reported to the security team relied on CREATE
RULE to convert the internally-built temp table to a view. That's not
possible since commit b23cd185fd, and I was not able to find a
different way to turn the SELECT on the temp table into code
execution, so as far as I know this is only exploitable in v15 and
below. That's a fiddly assumption though, so apply this patch to
master and all stable versions.
Thanks to Pedro Gallegos for the report.
Security: CVE-2023-5869
Reviewed-by: Noah Misch
An OS crash could leave PG_VERSION empty or missing. The same symptom
appeared in a backup by block device snapshot, taken after the next
checkpoint and before the OS flushes the PG_VERSION blocks. Device
snapshots are not a documented backup method, however. Back-patch to
v15, where commit 9c08aea6a3 introduced
STRATEGY=WAL_LOG and made it the default.
Discussion: https://postgr.es/m/20240130195003.0a.nmisch@google.com
Restoring a base backup taken in the middle of CreateDirAndVersionFile()
or write_relmap_file() would lose the function's effects. The symptom
was absence of the database directory, PG_VERSION file, or
pg_filenode.map. If missing the directory, recovery would fail. Either
missing file would not fail recovery but would render the new database
unusable. Fix CreateDirAndVersionFile() with the transam/README "action
first and then write a WAL entry" strategy. That has a side benefit of
moving filesystem mutations out of a critical section, reducing the ways
to PANIC. Fix the write_relmap_file() call with a lock acquisition, so
it interacts with checkpoints like non-CREATE DATABASE calls do.
Back-patch to v15, where commit 9c08aea6a3
introduced STRATEGY=WAL_LOG and made it the default.
Discussion: https://postgr.es/m/20240130195003.0a.nmisch@google.com
DST law changes in Ittoqqortoormiit, Greenland (America/Scoresbysund),
Kazakhstan (Asia/Almaty and Asia/Qostanay) and Palestine; as well as
updates for the Antarctic stations Casey and Vostok.
Historical corrections for Vietnam, Toronto, and Miquelon.
Further fallout from commit 6ee26c6a4b. To keep code in sync and avoid
issues on older releases with different package names, simply use the
unqualified name like many other places in our code.
The path we wish to reparameterize is not a standalone object:
in particular, it implicitly references baserestrictinfo clauses
in the associated RelOptInfo, and if it's a SampleScan path then
there is also the TableSampleClause in the RTE to worry about.
Both of those could contain lateral references to the join partner
relation, which would need to be modified to refer to its child.
Since we aren't doing that, affected queries can give wrong answers,
or odd failures such as "variable not found in subplan target list",
or executor crashes. But we can't just summarily modify those
expressions, because they are shared with other paths for the rel.
We'd break things if we modify them and then end up using some
non-partitioned-join path.
In HEAD, we plan to fix this by postponing reparameterization
until create_plan(), when we know that those other paths are
no longer of interest, and then adjusting those expressions along
with the ones in the path itself. That seems like too big a change
for stable branches however. In the back branches, let's just detect
whether any troublesome lateral references actually exist in those
expressions, and fail reparameterization if so. This will result in
not performing a partitioned join in such cases. Given the lack of
field complaints, nobody's likely to miss the optimization.
Report and patch by Richard Guo. Apply to 12-16 only, since
the intended fix for HEAD looks quite different. We're not quite
ready to push the HEAD fix, but with back-branch releases coming
up soon, it seems wise to get this stopgap fix in place there.
Discussion: https://postgr.es/m/CAMbWs496+N=UAjOc=rcD3P7B6oJe4rZw08e_TZRUsWbPxZW3Tw@mail.gmail.com
This impacts the statistics retrieved in transactions for the following
views when updating the value of stats_fetch_consistency, leading to
behaviors contrary to what is documented since 605994651b as an update
of this parameter should discard all statistics snapshot data:
- pg_stat_archiver
- pg_stat_bgwriter
- pg_stat_checkpointer
- pg_stat_io
- pg_stat_slru
- pg_stat_wal
For example, updating stats_fetch_consistency from "snapshot" to "cache"
in a transaction did not re-fetch any fresh data, using data cached from
the time when "snapshot" was in use.
Author: Shinya Kato
Discussion: https://postgr.es/m/d77fc5190d4dbe1738d77231488e768b@oss.nttdata.com
Backpatch-through: 15
This commit addresses a set of issues when changing token type mappings
in a text search configuration when using duplicated token names:
- ADD MAPPING would fail on insertion because of a constraint failure
after inserting the same mapping.
- ALTER MAPPING with an "overridden" configuration failed with "tuple
already updated by self" when the token mappings are removed.
- DROP MAPPING failed with "tuple already updated by self", like
previously, but in a different code path.
The code is refactored so the token names (with their numbers) are
handled as a List with unique members rather than an array with numbers,
ensuring that no duplicates mess up with the catalog inserts, updates
and deletes. The list is generated by getTokenTypes(), with the same
error handling as previously while duplicated tokens are discarded from
the list used to work on the catalogs.
Regression tests are expanded to cover much more ground for the cases
fixed by this commit, as there was no coverage for the code touched in
this commit. A bit more is done regarding the fact that a token name
not supported by a configuration's parser should result in an error even
if IF EXISTS is used in a DROP MAPPING clause. This is implied in the
code but there was no coverage for that, and it was very easy to miss.
These issues exist since at least their introduction in core with
140d4ebcb4, so backpatch all the way down.
Reported-by: Alexander Lakhin
Author: Tender Wang, Michael Paquier
Discussion: https://postgr.es/m/18310-1eb233c5908189c8@postgresql.org
Backpatch-through: 12
File::Find converts backslashes to slashes in the newer Perl versions.
See: 414f14df98
So, do the same conversion for Windows before comparing paths. To
support all Perl versions, always convert them on Windows regardless of
the Perl's version.
Author: Nazir Bilal Yavuz <byavuz81@gmail.com>
Backpatch to all live branches
We seem to have only documented a foreign key can reference the columns of
a primary key or unique constraint. Here we adjust the documentation
to mention columns in a non-partial unique index can be mentioned too.
The header comment for transformFkeyCheckAttrs() also didn't mention
unique indexes, so fix that too. In passing make that header comment
reflect reality in the various other aspects where it deviated from it.
Bug: 18295
Reported-by: Gilles PARC
Author: Laurenz Albe, David Rowley
Discussion: https://www.postgresql.org/message-id/18295-0ed0fac5c9f7b17b%40postgresql.org
Backpatch-through: 12
This function requires simd.h, which is a rather large dependency
for a widely-used header file like pg_wchar.h. Furthermore, there
is a report of a third-party tool that is struggling to use
pg_wchar.h due to its dependence on simd.h (presumably because
simd.h uses several intrinsics). Moving the function to the much
less popular ascii.h resolves these issues for now.
This commit is back-patched for the benefit of the aforementioned
third-party tool. The simd.h dependency was only added in v16,
but we've opted to back-patch to v15 so that is_valid_ascii() lives
in the same file for all versions where it exists. This could
break existing third-party code that uses the function, but we
couldn't find any examples of such code. It should be possible to
fix any code that this commit breaks by including ascii.h in the
file that uses is_valid_ascii().
Author: Jubilee Young
Reviewed-by: Tom Lane, John Naylor, Andres Freund, Eric Ridge
Discussion: https://postgr.es/m/CAPNHn3oKJJxMsYq%2BqLYzVJOFrUcOr4OF1EC-KtFT-qh8nOOOtQ%40mail.gmail.com
Backpatch-through: 15
libxml2 changed the required signature of error handler callbacks
to make the passed xmlError struct "const". This is causing build
failures on buildfarm member caiman, and no doubt will start showing
up in the field quite soon. Add a version check to adjust the
declaration of xml_errorHandler() according to LIBXML_VERSION.
2.12.x also produces deprecation warnings for contrib/xml2/xpath.c's
assignment to xmlLoadExtDtdDefaultValue. I see no good reason for
that to still be there, seeing that we disabled external DTDs (at a
lower level) years ago for security reasons. Let's just remove it.
Back-patch to all supported branches, since they might all get built
with newer libxml2 once it gets a bit more popular. (The back
branches produce another deprecation warning about xpath.c's use of
xmlSubstituteEntitiesDefault(). We ought to consider whether to
back-patch all or part of commit 65c5864d7 to silence that. It's
less urgent though, since it won't break the buildfarm.)
Discussion: https://postgr.es/m/1389505.1706382262@sss.pgh.pa.us
ginFinishSplit() expects the caller to hold an exclusive lock on the
buffer, but when finishing an earlier "leftover" incomplete split of
an internal page, the caller held a shared lock. That caused an
assertion failure in MarkBufferDirty(). Without assertions, it could
lead to corruption if two backends tried to complete the split at the
same time.
On master, add a test case using the new injection point facility.
Report and analysis by Fei Changhong. Backpatch the fix to all
supported versions.
Reviewed-by: Fei Changhong, Michael Paquier
Discussion: https://www.postgresql.org/message-id/tencent_A3CE810F59132D8E230475A5F0F7A08C8307@qq.com
In commit 272248a0c, we fixed the catalog lookup due to the wrong snapshot
for transactions and subtransactions during decoding. We failed to
consider the case where top-level xact is already marked as containing
catalog change but its subtransaction is not yet marked as containing
catalog change even though it contained such a change.
This can happen when during decoding, none of the WAL records from the
subtransaction was decoded and top-level xact contains a DDL.
We fix it by marking the transaction and all its subtransactions as
containing catalog changes if the top-level xact contains any catalog
change and it is present in the initial running xacts array.
This fix is required only for 14 and 15 because in prior branches we
already always mark the transaction and all its subtransactions as
containing catalog changes in the same case. In 16 and above, we preserve
the list of transaction IDs and sub-transaction IDs, that have modified
catalogs and are running during snapshot serialization, to the serialized
snapshot (see commit 7f13ac8123).
Author: Fei Changhong
Reviewed-by: Amit Kapila, Hayato Kuroda, Andy Fan
Discussion: https://postgr.es/m/18280-4c8060178cb41750@postgresql.org
Three LOG messages are added in the recovery code paths, providing
information that can be useful to track corruption issues depending on
the state of the cluster, telling that:
- Recovery has started from a backup_label.
- Recovery is restarting from a backup start LSN, without a
backup_label.
- Recovery has completed from a backup.
This was originally applied on HEAD as of 1d35f705e1, and there is
consensus that this can be useful for older versions. This applies
cleanly down to 15, so do it down to this version for now (older
versions have heavily refactored the WAL recovery paths, making the
change less straight-forward to do).
Author: Andres Freund
Reviewed-by: David Steele, Laurenz Albe, Michael Paquier
Discussion: https://postgr.es/m/20231117041811.vz4vgkthwjnwp2pp@awork3.anarazel.de
Backpatch-through: 15
We perform addition of the days field of an interval via
arithmetic on the Julian-date representation of the timestamp's date.
This step is subject to int32 overflow, and we also should not let
the Julian date become very negative, for fear of weird results from
j2date. (In the timestamptz case, allow a Julian date of -1 to pass,
since it might convert back to zero after timezone rotation.)
The additions of the months and microseconds fields could also
overflow, of course. However, I believe we need no additional
checks there; the existing range checks should catch such cases.
The difficulty here is that j2date's magic modular arithmetic could
produce something that looks like it's in-range.
Per bug #18313 from Christian Maurer. This has been wrong for
a long time, so back-patch to all supported branches.
Discussion: https://postgr.es/m/18313-64d2c8952d81e84b@postgresql.org
This command, when used to add a column on a parent table with a complex
inheritance tree, tried to update multiple times the same tuple in
pg_attribute for a child table when incrementing attinhcount, causing
failures with "tuple already updated by self" because of a missing
CommandCounterIncrement() between two updates.
This exists for a rather long time, so backpatch all the way down.
Reported-by: Alexander Lakhin
Author: Tender Wang
Reviewed-by: Richard Guo
Discussion: https://postgr.es/m/18297-b04cd83a55b51e35@postgresql.org
Backpatch-through: 12
When a pipeline is opened with \startpipeline and not closed, pgbench
will either error on the next transaction with a "already in pipeline
mode" error or successfully end if this was the last transaction --
despite not sending anything that was piped in the pipeline.
Make it an error to reach end of script is reached while there's an
open pipeline.
Backpatch to 14, where pgbench got support for pipelines.
Author: Anthonin Bonnefoy <anthonin.bonnefoy@datadoghq.com>
Reported-by: Michael Paquier <michael@paquier.xyz>
Discussion: https://postgr.es/m/Za4IObZkDjrO4TcS@paquier.xyz
plpgsql fails on new-style CREATE FUNCTION/PROCEDURE commands within
a routine or DO block, because make_execsql_stmt believes that a
semicolon token always terminates a SQL command. Now, that's actually
been wrong since the day it was written, because CREATE RULE has long
allowed multiple rule actions separated by semicolons. But there are
few enough people using multi-action rules that there was never an
attempt to fix it. New-style SQL functions, though, are popular.
psql has this same problem of "does this semicolon really terminate
the command?". It deals with CREATE RULE by counting parenthesis
nesting depth: a semicolon within parens doesn't end a command.
Commits e717a9a18 and 029c5ac03 created a similar heuristic to count
matching BEGIN/END pairs (but only within CREATEs, so as not to be
fooled by plain BEGIN). That's survived several releases now without
trouble reports, so let's just absorb those heuristics into plpgsql.
Per report from Samuel Dussault. Back-patch to v14 where new-style
SQL function syntax came in.
Discussion: https://postgr.es/m/YT2PR01MB88552C3E9AD40A6C038774A781722@YT2PR01MB8855.CANPRD01.PROD.OUTLOOK.COM
A REINDEX INDEX done on a partitioned index builds a list of the indexes
to work on before processing its partitions in individual transactions.
When combined with a DROP of the partitioned index, there was a window
where it was possible to see some unexpected "could not open relation
with OID", synonym of relation lookup error. The code was robust enough
to handle the case where the parent relation is missing, but not the
case where an index would be gone missing.
This is similar to 1d65416661.
Support for REINDEX on partitioned relations has been introduced in
a6642b3ae0, so backpatch down to 14.
Author: Fei Changhong
Discussion: https://postgr.es/m/tencent_6A52106095ACDE55333E3AD33F304C0C3909@qq.com
Backpatch-through: 14
try_index_open() is able to open an index if its relkind fits, except
that it would return NULL instead of generated an error if the relation
does not exist. This new routine will be used by an upcoming patch to
make REINDEX on partitioned relations more robust when an index in a
partition tree is dropped.
Extracted from a larger patch by the same author.
Author: Fei Changhong
Discussion: https://postgr.es/m/tencent_6A52106095ACDE55333E3AD33F304C0C3909@qq.com
Backpatch-through: 14
Until now LWLockDequeueSelf() sequentially searched the list of waiters to see
if the current proc is still is on the list of waiters, or has already been
removed. In extreme workloads, where the wait lists are very long, this leads
to a quadratic behavior. #backends iterating over a list #backends
long. Additionally, the likelihood of needing to call LWLockDequeueSelf() in
the first place also increases with the increased length of the wait queue, as
it becomes more likely that a lock is released while waiting for the wait list
lock, which is held for longer during lock release.
Due to the exponential back-off in perform_spin_delay() this is surprisingly
hard to detect. We should make that easier, e.g. by adding a wait event around
the pg_usleep() - but that's a separate patch.
The fix is simple - track whether a proc is currently waiting in the wait list
or already removed but waiting to be woken up in PGPROC->lwWaiting.
In some workloads with a lot of clients contending for a small number of
lwlocks (e.g. WALWriteLock), the fix can substantially increase throughput.
This has been originally fixed for 16~ with a4adc31f69 without a
backpatch, and we have heard complaints from users impacted by this
quadratic behavior in older versions as well.
Author: Andres Freund <andres@anarazel.de>
Reviewed-by: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
Discussion: https://postgr.es/m/20221027165914.2hofzp4cvutj6gin@awork3.anarazel.de
Discussion: https://postgr.es/m/CALj2ACXktNbG=K8Xi7PSqbofTZozavhaxjatVc14iYaLu4Maag@mail.gmail.com
Backpatch-through: 12
The comment was copy-pasted from the call to ProcSignalInit() in
AuxiliaryProcessMain(), which uses a similar scheme of having reserved
slots for aux processes after MaxBackends slots for backends. However,
ProcSignalInit() indexing starts from 1, whereas BackendStatusArray
starts from 0. The code is correct, but the comment was wrong.
Discussion: https://www.postgresql.org/message-id/f3ecd4cb-85ee-4e54-8278-5fabfb3a4ed0@iki.fi
Backpatch-through: v14
If configuring the newly created socket non-blocking fails we
error out and return INVALID_SOCKET, but the socket that had
been created wasn't closed. Fix by issuing closesocket in the
errorpath.
Backpatch to all supported branches.
Author: Ranier Vilela <ranier.vf@gmail.com>
Discussion: https://postgr.es/m/CAEudQApmU5CrKefH85VbNYE2y8H=-qqEJbg6RAPU65+vCe+89A@mail.gmail.com
Backpatch-through: v12
Commit b8ba7344e9 added in PQgetResult a derefence to a pointer
returned by pqPrepareAsyncResult(), before some other code that was
already testing that pointer for nullness. But since commit
618c16707a (in Postgres 15), pqPrepareAsyncResult() doesn't ever
return NULL (a statically-allocated result is returned if OOM). So in
branches 15 and up, we can remove the redundant pointer check with no
harm done.
However, in branch 14, pqPrepareAsyncResult() can indeed return NULL if
it runs out of memory. Fix things there by adding a null pointer check
before dereferencing the pointer. This should hint Coverity that the
preexisting check is not redundant but necessary.
Backpatch to 14, like b8ba7344e9.
Per Coverity.
When ExecBRUpdateTriggers switches to a new target tuple as a result
of the EvalPlanQual logic, it must form a new proposed update tuple.
Since commit 86dc90056, that tuple (the result of
ExecGetUpdateNewTuple) has been a virtual tuple that might contain
pointers to by-ref fields of the new target tuple (in "oldslot").
However, immediately after that we materialize oldslot, causing it to
drop its buffer pin, whereupon the by-ref pointers are unsafe to use.
This is a live bug only when the new target tuple is in a different
page than the original target tuple, since we do still hold a pin on
the original one. (Before 86dc90056, there was no bug because the
EPQ plantree would hold a pin on the new target tuple; but now that's
not assured.) To fix, forcibly materialize the new tuple before we
materialize oldslot. This costs nothing since we would have done that
shortly anyway.
The real-world impact of this is probably minimal. A visible failure
could occur if the new target tuple's buffer were recycled for some
other page in the short interval before we materialize newslot within
the trigger-calling loop; but that's quite unlikely given that we'd
just touched that page. There's a larger hazard that some other
process could prune and repack that page within the window. We have
lock on the new target tuple, but that wouldn't prevent it being moved
on the page.
Alexander Lakhin and Tom Lane, per bug #17798 from Alexander Lakhin.
Back-patch to v14 where 86dc90056 came in.
Discussion: https://postgr.es/m/17798-0907404928dcf0dd@postgresql.org
We've long had a policy that any toasted fields in a catalog tuple
should be pulled in-line before entering the tuple in a catalog cache.
However, that requires access to the catalog's toast table, and we'll
typically do AcceptInvalidationMessages while opening the toast table.
So it's possible that the catalog tuple is outdated by the time we
finish detoasting it. Since no cache entry exists yet, we can't
mark the entry stale during AcceptInvalidationMessages, and instead
we'll press forward and build an apparently-valid cache entry. The
upshot is that we have a race condition whereby an out-of-date entry
could be made in a backend's catalog cache, and persist there
indefinitely causing indeterminate misbehavior.
To fix, use the existing systable_recheck_tuple code to recheck
whether the catalog tuple is still up-to-date after we finish
detoasting it. If not, loop around and restart the process of
searching the catalog and constructing cache entries from the top.
The case is rare enough that this shouldn't create any meaningful
performance penalty, even in the SearchCatCacheList case where
we need to tear down and reconstruct the whole list.
Indeed, the case is so rare that AFAICT it doesn't occur during
our regression tests, and there doesn't seem to be any easy way
to build a test that would exercise it reliably. To allow
testing of the retry code paths, add logic (in USE_ASSERT_CHECKING
builds only) that randomly pretends that the recheck failed about
one time out of a thousand. This is enough to ensure that we'll
pass through the retry paths during most regression test runs.
By adding an extra level of looping, this commit creates a need
to reindent most of SearchCatCacheMiss and SearchCatCacheList.
I'll do that separately, to allow putting those changes in
.git-blame-ignore-revs.
Patch by me; thanks to Alexander Lakhin for having built a test
case to prove the bug is real, and to Xiaoran Wang for review.
Back-patch to all supported branches.
Discussion: https://postgr.es/m/1393953.1698353013@sss.pgh.pa.us
Discussion: https://postgr.es/m/CAGjhLkOoBEC9mLsnB42d3CO1vcMx71MLSEuigeABbQ8oRdA6gw@mail.gmail.com
This is similar to 9886744a36, to prevent the execution of other
programs due to autorun configurations which could influence the
postmaster startup.
This was originally applied on HEAD as of 83c75ac7fb69 without a
backpatch, but the patch has survived CI and buildfarm cycles. I have
checked that cmd /d exists down to Windows XP, which should make this
change work correctly in the oldest branches still supported.
Discussion: https://postgr.es/m/20230922.161551.320043332510268554.horikyota.ntt@gmail.com
Backpatch-through: 12
On Windows, cmd.exe is used to launch the postmaster process to ease its
redirection setup. However, cmd.exe may execute other programs at
startup due to autorun configurations, which could influence the
postmaster startup. This patch adds /D flag to the launcher cmd.exe
command line to disable autorun settings written in the registry.
This was originally applied on HEAD as of 9886744a36 without a
backpatch, but the patch has survived CI and buildfarm cycles. I have
checked that cmd /d exists down to Windows XP, which should make this
change work correctly in the oldest branches still supported.
Reported-by: Hayato Kuroda
Author: Kyotaro Horiguchi
Reviewed-by: Robert Haas, Michael Paquier
Discussion: https://postgr.es/m/20230922.161551.320043332510268554.horikyota.ntt@gmail.com
Backpatch-through: 12
The code for wrapping subquery output expressions in PlaceHolderVars
believed that if the expression already was a PlaceHolderVar, it was
never necessary to wrap that in another one. That's wrong if the
expression is underneath an outer join and involves a lateral
reference to outside that scope: failing to add an additional PHV
risks evaluating the expression at the wrong place and hence not
forcing it to null when the outer join should do so. This is an
oversight in commit 9e7e29c75, which added logic to forcibly wrap
lateral-reference Vars in PlaceHolderVars, but didn't see that the
adjacent case for PlaceHolderVars needed the same treatment.
The test case we have for this doesn't fail before 4be058fe9, but now
that I see the problem I wonder if it is possible to demonstrate
related errors before that. That's moot though, since all such
branches are out of support.
Per bug #18284 from Holger Reise. Back-patch to all supported
branches.
Discussion: https://postgr.es/m/18284-47505a20c23647f8@postgresql.org
Commit 9d9c02ccd, which added the notion of a "run condition" for
window functions, neglected to teach nodeFuncs.c to process the new
field. Remarkably, that doesn't seem to have had any ill effects
before we invented Var.varnullingrels, but now it can cause visible
failures in join-removal scenarios.
I have no faith that there's not reachable problems in v15 too,
so back-patch the code change to v15 where 9d9c02ccd came in.
The test case seems irrelevant to v15, though.
Per bug #18277 from Zuming Jiang. Diagnosis and patch by
Richard Guo.
Discussion: https://postgr.es/m/18277-089ead83b329a2fd@postgresql.org
During the calculations of the maximum for the number of buckets, take into
account that later we round that to the next power of 2.
Reported-by: Karen Talarico
Bug: #16925
Discussion: https://postgr.es/m/16925-ec96d83529d0d629%40postgresql.org
Author: Thomas Munro, Andrei Lepikhov, Alexander Korotkov
Reviewed-by: Alena Rybakina
Backpatch-through: 12
CheckPWChallengeAuth() would return STATUS_ERROR if the user does not
exist or has no password assigned, even if the client disconnected
without responding to the password challenge (as libpq often will,
for example). We should return STATUS_EOF in that case, and the
lower-level functions do, but this code level got it wrong since the
refactoring done in 7ac955b34. This breaks the intent of not logging
anything for EOF cases (cf. comments in auth_failed()) and might
also confuse users of ClientAuthentication_hook.
Per report from Liu Lang. Back-patch to all supported versions.
Discussion: https://postgr.es/m/b725238c-539d-cb09-2bff-b5e6cb2c069c@esgyn.cn
If the underlying table isn't being dumped, it's useless to dump
an extended statistics object; it'll just cause errors at restore.
We have always applied similar policies to, say, indexes.
(When and if we get cross-table stats objects, it might be profitable
to think a little harder about what to do with them. But for now
there seems no point in considering a stats object as anything but
an appendage of its table.)
Rian McGuire and Tom Lane, per report from Rian McGuire.
Back-patch to supported branches.
Discussion: https://postgr.es/m/7075d3aa-3f05-44a5-b68f-47dc6a8a0550@buildkite.com
set_config_option() bails out early if it detects that the option to
be set is PGC_BACKEND or PGC_SU_BACKEND class and we're reading the
config file in a postmaster child; we don't want to apply any new
value in such a case. That's fine as far as it goes, but it fails
to consider the requirements of the pg_file_settings view: for that,
we need to check validity of the value even though we have no
intention to apply it. Because we didn't, even very silly values
for affected GUCs would be reported as valid by the view. There
are only half a dozen such GUCs, which perhaps explains why this
got overlooked for so long.
Fix by continuing when changeVal is false; this parallels the logic
in some other early-exit paths.
Also, the check added by commit 924bcf4f1 to prevent GUC changes in
parallel workers seems a few bricks shy of a load: it's evidently
assuming that ereport(elevel, ...) won't return. Make sure we
bail out if it does. The lack of trouble reports suggests that
this is only a latent bug, i.e. parallel workers don't actually
reach here with elevel < ERROR. (Per the code coverage report,
we never reach here at all in the regression suite.) But we clearly
don't want to risk proceeding if that does happen.
Per report from Rıdvan Korkmaz. These are ancient bugs, so back-patch
to all supported branches.
Discussion: https://postgr.es/m/2089235.1703617353@sss.pgh.pa.us
Like commit 388e80132, use "#pragma GCC system_header" to silence
warnings appearing within the Python headers, since newer Python
versions no longer worry about some restrictions we still use like
-Wdeclaration-after-statement.
This patch improves on 388e80132 by inventing a separate wrapper
header file, allowing the pragma to be tightly scoped to just
the Python headers and not other stuff we have laying about in
plpython.h. I applied the same technique to plperl for the same
reason: the original patch suppressed warnings for a good deal
of our own code, not only the Perl headers.
Like the previous commit, back-patch to supported branches.
Peter Eisentraut and Tom Lane
Discussion: https://postgr.es/m/ae523163-6d2a-4b81-a875-832e48dec502@eisentraut.org
This is necessary when spgcanreturn() is invoked on a partitioned
index, and the failure might be reachable in other scenarios as
well. The rest of what spgGetCache() does is perfectly sensible
for a partitioned index, so we should allow it to go through.
I think the main takeaway from this is that we lack sufficient test
coverage for non-btree partitioned indexes. Therefore, I added
simple test cases for brin and gin as well as spgist (hash and
gist AMs were covered already in indexing.sql).
Per bug #18256 from Alexander Lakhin. Although the known test case
only fails since v16 (3c569049b), I've got no faith at all that there
aren't other ways to reach this problem; so back-patch to all
supported branches.
Discussion: https://postgr.es/m/18256-0b0e1b6e4a620f1b@postgresql.org
Fix a bug during MERGE if a cross-partition update is attempted on a
partitioned table with a BEFORE DELETE ROW trigger that returns NULL,
to prevent the update. This would cause an error to be thrown, or an
assert failure in an assert-enabled build.
This was an oversight in 9321c79c86, which failed to properly
distinguish a DELETE prevented by a trigger from one prevented by a
concurrent update. Fix by having ExecDelete() return the TM_Result
status to ExecCrossPartitionUpdate(), so that it can distinguish the
two cases, and make ExecCrossPartitionUpdate() return the TM_Result
status to ExecUpdateAct(), so that it can return the correct status
from a concurrent update.
In addition, ensure that the command tag is correctly updated by
having ExecMergeMatched() pass canSetTag to ExecUpdateAct(), rather
than passing false, so that it updates the command tag if it does a
cross-partition update, making this code path in ExecMergeMatched()
consistent with ExecUpdate().
Per bug #18238 from Alexander Lakhin. Back-patch to v15, where MERGE
was introduced.
Dean Rasheed, reviewed by Richard Guo and Jian He.
Discussion: https://postgr.es/m/18238-2f2bdc7f720180b9%40postgresql.org
In v16 and up (since commit afbfc0298), large object ownership
checking has been broken because object_ownercheck() didn't take care
of the discrepancy between our object-address representation of large
objects (classId == LargeObjectRelationId) and the catalog where their
ownership info is actually stored (LargeObjectMetadataRelationId).
This resulted in failures such as "unrecognized class ID: 2613"
when trying to update blob properties as a non-superuser.
Poking around for related bugs, I found that AlterObjectOwner_internal
would pass the wrong classId to the PostAlterHook in the no-op code
path where the large object already has the desired owner. Also,
recordExtObjInitPriv checked for the wrong classId; that bug is only
latent because the stanza is dead code anyway, but as long as we're
carrying it around it should be less wrong. These bugs are quite old.
In HEAD, we can reduce the scope for future bugs of this ilk by
changing AlterObjectOwner_internal's API to let the translation happen
inside that function, rather than requiring callers to know about it.
A more bulletproof fix, perhaps, would be to start using
LargeObjectMetadataRelationId as the dependency and object-address
classId for blobs. However that has substantial risk of breaking
third-party code; even within our own code, it'd create hassles
for pg_dump which would have to cope with a version-dependent
representation. For now, keep the status quo.
Discussion: https://postgr.es/m/2650449.1702497209@sss.pgh.pa.us
This fixes an error introduced by efb0ef909f, that changed the
description of this field to "shared/local" while these I/O timings
relate to shared buffers. This information is available when
track_io_timing is enabled. Note that HEAD has added new counters for
local buffers in 295c36c0c1, so there is no need to touch it. The
description is updated to "shared" to be compatible with HEAD.
Per discussion with Nazir Bilal Yavuz and Hubert Depesz Lubaczewski,
whose EXPLAIN analyzer tool was not actually able to parse the previous
term because of the slash character.
Discussion: https://postgr.es/m/ZTCTiUqm_H3iBihl@paquier.xyz
Backpatch-through: 15
Dead tuples are ignored and are not marked as dead during recovery, as
it can lead to MVCC issues on a standby because its xmin may not match
with the primary. This information is tracked by a field called
"xactStartedInRecovery" in the transaction state data, switched on when
starting a transaction in recovery.
Unfortunately, this information was not correctly tracked when starting
a subtransaction, because the transaction state used for the
subtransaction did not update "xactStartedInRecovery" based on the state
of its parent. This would cause index scans done in subtransactions to
return inconsistent data, depending on how the xmin of the primary
and/or the standby evolved.
This is broken since the introduction of hot standby in efc16ea520, so
backpatch all the way down.
Author: Fei Changhong
Reviewed-by: Kyotaro Horiguchi
Discussion: https://postgr.es/m/tencent_C4D907A5093C071A029712E73B43C6512706@qq.com
Backpatch-through: 12
OpenSSL will sometimes return SSL_ERROR_SYSCALL without having set
errno; this is apparently a reflection of recv(2)'s habit of not
setting errno when reporting EOF. Ensure that we treat such cases
the same as read EOF. Previously, we'd frequently report them like
"could not accept SSL connection: Success" which is confusing, or
worse report them with an unrelated errno left over from some
previous syscall.
To fix, ensure that errno is zeroed immediately before the call,
and report its value only when it's not zero afterwards; otherwise
report EOF.
For consistency, I've applied the same coding pattern in libpq's
pqsecure_raw_read(). Bare recv(2) shouldn't really return -1 without
setting errno, but in case it does we might as well cope.
Per report from Andres Freund. Back-patch to all supported versions.
Discussion: https://postgr.es/m/20231208181451.deqnflwxqoehhxpe@awork3.anarazel.de
The apply worker needs to update the state of the subscription tables to
'READY' during the synchronization phase which requires locking the
corresponding subscription. The apply worker also waits for the
subscription tables to reach the 'SYNCDONE' state after holding the locks
on the subscription and the wait is done using WaitLatch. The 'SYNCDONE'
state is changed by tablesync workers again by locking the corresponding
subscription. Both the state updates use AccessShareLock mode to lock the
subscription, so they can't block each other. However, a backend can
simultaneously try to acquire a lock on the same subscription using
AccessExclusiveLock mode to alter the subscription. Now, the backend's
wait on a lock can sneak in between the apply worker and table sync worker
causing deadlock.
In other words, apply_worker waits for tablesync worker which waits for
backend, and backend waits for apply worker. This is not detected by the
deadlock detector because apply worker uses WaitLatch.
The fix is to release existing locks in apply worker before it starts to
wait for tablesync worker to change the state.
Reported-by: Tomas Vondra
Author: Shlok Kyal
Reviewed-by: Amit Kapila, Peter Smith
Backpatch-through: 12
Discussion: https://postgr.es/m/d291bb50-12c4-e8af-2af2-7bb9bb4d8e3e@enterprisedb.com
While checking if a record could fit in the circular WAL decoding
buffer, the coding from commit 3f1ce973 used arithmetic that could
overflow. 64 bit systems were unaffected for various technical reasons,
which probably explains the lack of problem reports. Likewise for 32
bit systems running known 32 bit kernels. The systems at risk of
problems appear to be 32 bit processes running on 64 bit kernels, with
unlucky placement in memory.
Per complaint from GCC -fsanitize=undefined -m32, while testing
variations of 039_end_of_wal.pl.
Back-patch to 15.
Reviewed-by: Nathan Bossart <nathandbossart@gmail.com>
Reviewed-by: Robert Haas <robertmhaas@gmail.com>
Discussion: https://postgr.es/m/CA%2BhUKGKH0oRPOX7DhiQ_b51sM8HqcPp2J3WA-Oen%3DdXog%2BAGGQ%40mail.gmail.com
During a pg_upgrade test using an old dump, all references to the old
regress shared library path (so, dylib or dll) are updated to point to
the library path used by the new build, to ensure a consistent
comparison between the old and new dumps.
The test previously relied on a hardcoded value of "src/test/regress/"
to build the new path value, which would point to an incorrect location
for the meson and vpath builds. This is replaced by REGRESS_SHLIB, able
to point to the correct location of the regress shared library.
Author: Alexander Lakhin
Discussion: https://postgr.es/m/a628d8ad-a08a-2eab-4ca9-641bc82d3193@gmail.com
Backpatch-through: 15
This commit removes the restriction that would not apply filters to the
dumps used for comparison in the TAP test of pg_upgrade when using the
same base version for the old and new nodes.
The previous logic would fail on Windows if loading a dump while using
the same set of binaries for the old and new nodes, as the library
dependencies updated in the old dump would append CRLFs to the dump
file as it is treated as a text file. The dump filtering logic replaces
all CRLFs (\r\n) by LFs (\n), which is able to prevent this issue.
When the old and new versions of the binaries are the same,
AdjustUpgrade removes all blank lines, removes version-based comments
generated by pg_dump and replaces CRLFs by LFs.
Reported-by: Alexander Lakhin
Discussion: https://postgr.es/m/60d434b9-53d9-9ea1-819b-efebdcf44e41@gmail.com
Backpatch-through: 15
Commit 5a991ef869 accidentally reversed the order of the tuples
and fields parameters, making the error message incorrectly refer
to 3 tuples with 1 field when IDENTIFY_SYSTEM returns 1 tuple and
3 or 4 fields. Fix by changing the order of the parameters. This
also adds a comment describing why we check for < 3 when postgres
since 9.4 has been sending 4 fields.
Backpatch all the way since the bug is almost a decade old.
Author: Tomonari Katsumata <t.katsumata1122@gmail.com>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Bug: #18224
Backpatch-through: v12
The logic to keep the libpq command queue in sync with queries that have
been processed had a bug when errors were returned for reasons other
than problems in queries -- for example, when a connection is lost. We
incorrectly consumed an element from the command queue every time, but
this is wrong and can lead to the queue becoming empty ahead of time,
leading to later malfunction: PQgetResult would return nothing,
potentially causing the calling application to enter a busy loop.
Fix by making the SYNC queue element a barrier that can only be consumed
when a SYNC message is received.
Backpatch to 14.
Reported by: Иван Трофимов (Ivan Trofimov) <i.trofimow@yandex.ru>
Discussion: https://postgr.es/m/17948-fcace7557e449957@postgresql.org
It draws an unnecessary error in builds compiled without thread support.
Added by commit 038f586d5f, which was backpatched to 14; though in
branch master we no longer support such builds, there's no reason to
have this there, so remove it in all branches since 14.
Reported-by: Michael Paquier <michael@paquier.xyz>
Discussion: https://postgr.es/m/ZW2G9Ix4nBKLcSSO@paquier.xyz
When creating a partitioned index, the partition key must be a subset
of the index's columns. But this currently doesn't check that the
collations between the partition key and the index definition match.
So you can construct a unique index that fails to enforce uniqueness.
(This would most likely involve a nondeterministic collation, so it
would have to be crafted explicitly and is not something that would
just happen by accident.)
This patch adds the required collation check. As a result, any
previously allowed unique index that has a collation mismatch would no
longer be allowed to be created.
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://www.postgresql.org/message-id/flat/3327cb54-f7f1-413b-8fdb-7a9dceebb938%40eisentraut.org
We should have done it this way all along, but we accidentally got
away with using the wrong BIO field up until OpenSSL 3.2. There,
the library's BIO routines that we rely on use the "data" field
for their own purposes, and our conflicting use causes assorted
weird behaviors up to and including core dumps when SSL connections
are attempted. Switch to using the approved field for the purpose,
i.e. app_data.
While at it, remove our configure probes for BIO_get_data as well
as the fallback implementation. BIO_{get,set}_app_data have been
there since long before any OpenSSL version that we still support,
even in the back branches.
Also, update src/test/ssl/t/001_ssltests.pl to allow for a minor
change in an error message spelling that evidently came in with 3.2.
Tristan Partin and Bo Andreson. Back-patch to all supported branches.
Discussion: https://postgr.es/m/CAN55FZ1eDDYsYaL7mv+oSLUij2h_u6hvD4Qmv-7PK7jkji0uyQ@mail.gmail.com
If the tuple being updated is not visible to the crosscheck snapshot,
we return TM_Updated but the assertions would not hold in that case.
Move them to before the cross-check.
Fixes bug #17893. Backpatch to all supported versions.
Author: Alexander Lakhin
Backpatch-through: 12
Discussion: https://www.postgresql.org/message-id/17893-35847009eec517b5%40postgresql.org
52e4f0cd4 introduced a bug in pgoutput in which missing values in tuples
were incorrectly filled in with NULL. The problem was the use of
CreateTupleDescCopy where CreateTupleDescCopyConstr was required, as the
former drops the constraints in the tuple description (specifically, the
default value constraint) on the floor.
The bug could result in incorrectness when a table replicated via
`REPLICA IDENTITY FULL` underwent a schema change that added a column
with a default value. The problem is that in such cases updates fill NULL
values in old tuples for missing columns for default values. Then on the
subscriber, we failed to find a matching tuple and missed updating the
required row.
Author: Nikhil Benesch
Reviewed-by: Hou Zhijie, Amit Kapila
Backpatch-through: 15
Discussion: http://postgr.es/m/CAPWqQZTEpZQamYsGMn6ZDRvVywwpVPiKH6OY4KSgA+NmeqFNzA@mail.gmail.com
The libpq code in charge of creating per-connection SSL objects was
prone to a race condition when loading the custom BIO methods needed by
my_SSL_set_fd(). As BIO methods are stored as a static variable, the
initialization of a connection could fail because it could be possible
to have one thread refer to my_bio_methods while it is being manipulated
by a second concurrent thread.
This error has been introduced by 8bb14cdd33, that has removed
ssl_config_mutex around the call of my_SSL_set_fd(), that itself sets
the custom BIO methods used in libpq. Like previously, the BIO method
initialization is now protected by the existing ssl_config_mutex, itself
initialized earlier for WIN32.
While on it, document that my_bio_methods is protected by
ssl_config_mutex, as this can be easy to miss.
Reported-by: Willi Mann
Author: Willi Mann, Michael Paquier
Discussion: https://postgr.es/m/e77abc4c-4d03-4058-a9d7-ef0035657e04@celonis.com
Backpatch-through: 12
When using GSSAPI encryption in non-blocking mode, libpq sometimes
failed with "GSSAPI caller failed to retransmit all data needing
to be retried". The cause is that pqPutMsgEnd rounds its transmit
request down to an even multiple of 8K, and sometimes that can lead
to not requesting a write of data that was requested to be written
(but reported as not written) earlier. That can upset pg_GSS_write's
logic for dealing with not-yet-written data, since it's possible
the data in question had already been incorporated into an encrypted
packet that we weren't able to send during the previous call.
We could fix this with a one-or-two-line hack to disable pqPutMsgEnd's
round-down behavior, but that seems like making the caller work around
a behavior that pg_GSS_write shouldn't expose in this way. Instead,
adjust pg_GSS_write to never report a partial write: it either
reports a complete write, or reflects the failure of the lower-level
pqsecure_raw_write call. The requirement still exists for the caller
to present at least as much data as on the previous call, but with
the caller-visible write start point not moving there is no temptation
for it to present less. We lose some ability to reclaim buffer space
early, but I doubt that that will make much difference in practice.
This also gets rid of a rather dubious assumption that "any
interesting failure condition (from pqsecure_raw_write) will recur
on the next try". We've not seen failure reports traceable to that,
but I've never trusted it particularly and am glad to remove it.
Make the same adjustments to the equivalent backend routine
be_gssapi_write(). It is probable that there's no bug on the backend
side, since we don't have a notion of nonblock mode there; but we
should keep the logic the same to ease future maintenance.
Per bug #18210 from Lars Kanis. Back-patch to all supported branches.
Discussion: https://postgr.es/m/18210-4c6d0b14627f2eb8@postgresql.org
If an error is thrown after calling CreateWaitEventSet(), the memory
of a WaitEventSet is free'd as it's allocated in the short-lived
memory context, but the file descriptor (on epoll- or kqueue-based
systems) or handles (on Windows) that it contains are leaked.
Use PG_TRY-FINALLY to ensure it gets freed. (On master, I will apply a
better fix, using ResourceOwners to track the WaitEventSet, but that's
not backpatchable.)
The added test doesn't check for leaking resources, so it passed even
before this commit. But at least it covers the code path.
In the passing, fix misleading comment on what the 'nevents' argument
to WaitEventSetWait means.
Report by Alexander Lakhin, analysis and suggestion for the fix by Tom
Lane. Fixes bug #17828. Backpatch to v14 where async execution was
introduced, but master gets a different fix.
Discussion: https://www.postgresql.org/message-id/17828-122da8cba23236be@postgresql.org
Discussion: https://www.postgresql.org/message-id/472235.1678387869@sss.pgh.pa.us
The copy command formed for initial sync was using parenthesis for tables
with no columns leading to syntax error. This patch avoids adding
parenthesis for such tables.
Reported-by: Justin G
Author: Vignesh C
Reviewed-by: Peter Smith, Amit Kapila
Backpatch-through: 15
Discussion: http://postgr.es/m/18203-df37fe354b626670@postgresql.org
As written, the query checked for an access method of type 's', which is
not an AM type supported in the core code.
Error introduced by 8586bf7ed8. As this query is not checking what it
should, backpatch all the way down.
Reviewed-by: Aleksander Alekseev
Discussion: https://postgr.es/m/ZVxJkAJrKbfHETiy@paquier.xyz
Backpatch-through: 12
The DROP STATISTICS code failed to properly lock the table, leading to
ERROR: tuple concurrently deleted
when executed concurrently with ANALYZE.
Fixed by modifying RemoveStatisticsById() to acquire the same lock as
ANALYZE. This function is called only by DROP STATISTICS, as ANALYZE
calls RemoveStatisticsDataById() directly.
Reported by Justin Pryzby, fix by me. Backpatch through 12. The code was
like this since it was introduced in 10, but older releases are EOL.
Reported-by: Justin Pryzby
Reviewed-by: Tom Lane
Backpatch-through: 12
Discussion: https://postgr.es/m/ZUuk-8CfbYeq6g_u@pryzbyj2023
Commits 146604ec43 and a898b409f6 added overflow checks to
interval_mul(), but not to interval_div(), which contains almost
identical code, and so is susceptible to the same kinds of
overflows. In addition, those checks did not catch all possible
overflow conditions.
Add additional checks to the "cascade down" code in interval_mul(),
and copy all the overflow checks over to the corresponding code in
interval_div(), so that they both generate "interval out of range"
errors, rather than returning bogus results.
Given that these errors are relatively easy to hit, back-patch to all
supported branches.
Per bug #18200 from Alexander Lakhin, and subsequent investigation.
Discussion: https://postgr.es/m/18200-5ea288c7b2d504b1%40postgresql.org
When performing inlining LLVM unfortunately "leaks" types (the
types survive and are usable, but a new round of inlining will
recreate new structurally equivalent types). This accumulation
will over time amount to a memory leak which for some queries
can be large enough to trigger the OOM process killer.
To avoid accumulation of types, all IR related data is stored
in an LLVMContextRef which is dropped and recreated in order
to release all types. Dropping and recreating incurs overhead,
so it will be done only after 100 queries. This is a heuristic
which might be revisited, but until we can get the size of the
context from LLVM we are flying a bit blind.
This issue has been reported several times, there may be more
references to it in the archives on top of the threads linked
below.
This is a backpatch of 9dce22033d to all supported branches.
Reported-By: Justin Pryzby <pryzby@telsasoft.com>
Reported-By: Kurt Roeckx <kurt@roeckx.be>
Reported-By: Jaime Casanova <jcasanov@systemguards.com.ec>
Reported-By: Lauri Laanmets <pcspets@gmail.com>
Author: Andres Freund and Daniel Gustafsson
Discussion: https://postgr.es/m/7acc8678-df5f-4923-9cf6-e843131ae89d@www.fastmail.com
Discussion: https://postgr.es/m/20201218235607.GC30237@telsasoft.com
Discussion: https://postgr.es/m/CAPH-tTxLf44s3CvUUtQpkDr1D8Hxqc2NGDzGXS1ODsfiJ6WSqA@mail.gmail.com
Backpatch-through: v12
contain_mutable_functions and contain_volatile_functions give
reliable answers only after expression preprocessing (specifically
eval_const_expressions). Some places understand this, but some did
not get the memo --- which is not entirely their fault, because the
problem is documented only in places far away from those functions.
Introduce wrapper functions that allow doing the right thing easily,
and add commentary in hopes of preventing future mistakes from
copy-and-paste of code that's only conditionally safe.
Two actual bugs of this ilk are fixed here. We failed to preprocess
column GENERATED expressions before checking mutability, so that the
code could fail to detect the use of a volatile function
default-argument expression, or it could reject a polymorphic function
that is actually immutable on the datatype of interest. Likewise,
column DEFAULT expressions weren't preprocessed before determining if
it's safe to apply the attmissingval mechanism. A false negative
would just result in an unnecessary table rewrite, but a false
positive could allow the attmissingval mechanism to be used in a case
where it should not be, resulting in unexpected initial values in a
new column.
In passing, re-order the steps in ComputePartitionAttrs so that its
checks for invalid column references are done before applying
expression_planner, rather than after. The previous coding would
not complain if a partition expression contains a disallowed column
reference that gets optimized away by constant folding, which seems
to me to be a behavior we do not want.
Per bug #18097 from Jim Keener. Back-patch to all supported versions.
Discussion: https://postgr.es/m/18097-ebb179674f22932f@postgresql.org
The fallback implementation of pg_atomic_test_set_flag() that uses
atomic-exchange gives pg_atomic_exchange_u32_impl() an extra
argument. This issue has been present since the introduction of
the atomics API in commit b64d92f1a5.
Reviewed-by: Andres Freund
Discussion: https://postgr.es/m/20231114035439.GA1809032%40nathanxps13
Backpatch-through: 12
checkExtensionMembership() set the DUMP_COMPONENT_SECLABEL and
DUMP_COMPONENT_POLICY flags for extension member objects, even though
we lack any infrastructure for tracking extensions' initial settings
of these properties. This is not OK. The result was that a dump
would always include commands to set these properties for extension
objects that have them, with at least three negative consequences:
1. The restoring user might not have privilege to set these properties
on these objects.
2. The properties might be incorrect/irrelevant for the version of the
extension that's installed in the destination database.
3. The dump itself might fail, in the case of RLS properties attached
to extension tables that the dumping user lacks privilege to LOCK.
(That's because we must get at least AccessShareLock to ensure that
we don't fail while trying to decompile the RLS expressions.)
When and if somebody cares to invent initial-state infrastructure for
extensions' RLS policies and security labels, we could think about
finding another way around problem #3. But in the absence of such
infrastructure, this whole thing is just wrong and we shouldn't do it.
(Note: this applies only to ordinary dumps; binary-upgrade dumps
still dump and restore extension member objects separately, with
all properties.)
Tom Lane and Jacob Champion. Back-patch to all supported branches.
Discussion: https://postgr.es/m/00d46a48-3324-d9a0-49bf-e7f0f11d1038@timescale.com
It's clearly stated in the comments that ginFindParents() must keep
the pin on the index's root page that's associated with the topmost
GinBtreeStack item. However, the code path for the case that the
desired downlink has been pushed down to the next index level
ignored this proviso, and would release the pin anyway if we were
still examining the root level. That led to an assertion failure
or "buffer NNNN is not owned by resource owner" error later, when
we try to release the pin again at the end of the insertion.
This is quite hard to reproduce, since it can only happen if an
index root page split occurs concurrently with our own insertion.
Thanks to Jeff Janes for finding a test case that triggers it
often enough to allow investigation.
This has been there since the beginning of GIN, so back-patch
to all supported branches.
Discussion: https://postgr.es/m/CAMkU=1yCAKtv86dMrD__Ja-7KzjE=uMeKX8y__cx5W-OEWy2ow@mail.gmail.com
Commit b7eda3e0e moved XidInMVCCSnapshot() from tqual.c into snapmgr.c,
but follow-up commit c91560def incorrectly updated this reference. We
could fix it, but as pointed out by Daniel Gustafsson, 1) the reader can
easily find the file that contains the definition of that function, e.g.
by grepping, and 2) this kind of reference is prone to going stale; so
let's just remove it.
Back-patch to all supported branches.
Reviewed by Daniel Gustafsson.
Discussion: https://postgr.es/m/CAPmGK145VdKkPBLWS2urwhgsfidbSexwY-9zCL6xSUJH%2BBTUUg%40mail.gmail.com
When executing a MERGE UPDATE action, if the UPDATE is turned into a
cross-partition DELETE then INSERT, do not attempt to invoke AFTER
UPDATE ROW triggers, or any of the other post-update actions in
ExecUpdateEpilogue().
For consistency with a plain UPDATE command, such triggers should not
be fired (and typically fail anyway), and similarly, other post-update
actions, such as WCO/RLS checks should not be executed, and might also
lead to unexpected failures.
Therefore, as with ExecUpdate(), make ExecMergeMatched() return
immediately if ExecUpdateAct() reports that a cross-partition update
was done, to be sure that no further processing is done for that
tuple.
Back-patch to v15, where MERGE was introduced.
Discussion: https://postgr.es/m/CAEZATCWjBgagyNZs02vgDF0DvASYj-iHTFtXG2-nP3orZhmtcw%40mail.gmail.com
When computing "0 - INT64_MIN", most platforms would report an
overflow error, which is correct. However, platforms without integer
overflow builtins or 128-bit integers would fail to spot the overflow,
and incorrectly return INT64_MIN.
Back-patch to all supported branches.
Patch be me. Thanks to Jian He for initial investigation, and Laurenz
Albe and Tom Lane for review.
Discussion: https://postgr.es/m/CAEZATCUNK-AZSD0jVdgkk0N%3DNcAXBWeAEX-QU9AnJPensikmdQ%40mail.gmail.com
When PQsendFlushRequest() was added by commit 69cf1d5429, we argued
against adding a PQflush() call in it[1]. This is still the right
decision: if the user wants a flush to occur, they can just call that.
However, we failed to realize that the message bytes could still be
given to the kernel for transmitting when this can be made without
blocking. That's what pqPipelineFlush() does, and it is done for every
single other message type sent by libpq, so do that.
(When the socket is in blocking mode this may indeed block, but that's
what all the other libpq message-sending routines do, too.)
[1] https://www.postgresql.org/message-id/202106252352.5ca4byasfun5%40alvherre.pgsql
Author: Jelte Fennema-Nio <postgres@jeltef.nl>
Discussion: https://postgr.es/m/CAGECzQTxZRevRWkKodE-SnJk1Yfm4eKT+8E4Cyq3MJ9YKTnNew@mail.gmail.com
false_positive_rate is a parameter that can be set with the bloom
opclass in BRIN, and setting it to a value of exactly 0.25 would trigger
an assertion in the first INSERT done on the index with value set.
The assertion changed here relied on BLOOM_{MIN|MAX}_FALSE_POSITIVE_RATE
that are somewhat arbitrary values, and specifying an out-of-range value
would also trigger a failure when defining such an index. So, as-is,
the assertion was just doubling on the min-max check of the reloption.
This is now enlarged to check that it is a correct percentage value,
instead, based on a suggestion by Tom Lane.
Author: Alexander Lakhin
Reviewed-by: Tom Lane, Shihao Zhong
Discussion: https://postgr.es/m/17969-a6c54de48026d694@postgresql.org
Backpatch-through: 14
array_set_element() and related functions allow an array to be
enlarged by assigning to subscripts outside the current array bounds.
While these places were careful to check that the new bounds are
allowable, they neglected to consider the risk of integer overflow
in computing the new bounds. In edge cases, we could compute new
bounds that are invalid but get past the subsequent checks,
allowing bad things to happen. Memory stomps that are potentially
exploitable for arbitrary code execution are possible, and so is
disclosure of server memory.
To fix, perform the hazardous computations using overflow-detecting
arithmetic routines, which fortunately exist in all still-supported
branches.
The test cases added for this generate (after patching) errors that
mention the value of MaxArraySize, which is platform-dependent.
Rather than introduce multiple expected-files, use psql's VERBOSITY
parameter to suppress the printing of the message text. v11 psql
lacks that parameter, so omit the tests in that branch.
Our thanks to Pedro Gallegos for reporting this problem.
Security: CVE-2023-5869
transformAggregateCall() captures the datatypes of the aggregate's
arguments immediately to construct the Aggref.aggargtypes list.
This seems reasonable because the arguments have already been
transformed --- but there is an edge case where they haven't been.
Specifically, if we have an unknown-type literal in an ANY argument
position, nothing will have been done with it earlier. But if we
also have DISTINCT, then addTargetToGroupList() converts the literal
to "text" type, resulting in the aggargtypes list not matching the
actual runtime type of the argument. The end result is that the
aggregate tries to interpret a "text" value as being of type
"unknown", that is a zero-terminated C string. If the text value
contains no zero bytes, this could result in disclosure of server
memory following the text literal value.
To fix, move the collection of the aggargtypes list to the end
of transformAggregateCall(), after DISTINCT has been handled.
This requires slightly more code, but not a great deal.
Our thanks to Jingzhou Fu for reporting this problem.
Security: CVE-2023-5868
It was always false in single-user mode, in autovacuum workers, and in
background workers. This had no specifically-identified security
consequences, but non-core code or future work might make it
security-relevant. Back-patch to v11 (all supported versions).
Jelte Fennema-Nio. Reported by Jelte Fennema-Nio.
Documentation says it cannot signal "a backend owned by a superuser".
On the contrary, it could signal background workers, including the
logical replication launcher. It could signal autovacuum workers and
the autovacuum launcher. Block all that. Signaling autovacuum workers
and those two launchers doesn't stall progress beyond what one could
achieve other ways. If a cluster uses a non-core extension with a
background worker that does not auto-restart, this could create a denial
of service with respect to that background worker. A background worker
with bugs in its code for responding to terminations or cancellations
could experience those bugs at a time the pg_signal_backend member
chooses. Back-patch to v11 (all supported versions).
Reviewed by Jelte Fennema-Nio. Reported by Hemanth Sandrana and
Mahendrakar Srinivasarao.
Security: CVE-2023-5870
Tom Lane
Amit Kapila
Alvaro Herrera
David Rowley
Michael Paquier
Heikki Linnakangas
Daniel Gustafsson
Alexander Korotkov
Peter Eisentraut
Noah Misch
Andrew Dunstan
Nathan Bossart
Thomas Munro
Andres Freund
Dean Rasheed
Tomas Vondra
Etsuro Fujita